UFlow

Moderator: Mike Everman

camino75080
Posts: 23
Joined: Wed Aug 04, 2004 8:32 am
Antipspambot question: 0
Location: Richardson, Tx
Contact:

UFlow

Could someone please explain how to use Uflow to simulate a pulsejet? I don't understand ant of the values exept the length snd daimeter. Thanks for any help.
A good friend will bail you out of jail, a great friend will be sitting in the cell with you laughing about how great it was, while you wait for a good friend.

pezman
Posts: 613
Joined: Fri Oct 08, 2004 4:13 am
Antipspambot question: 0
Location: USA

Re: UFlow

I have a similar question.

I was working on a valveless model -- basically the "ping-pong" design the Bruno Olgerec proposed for high-compression valveless pulse jets-- and simulated the initial conditions by splitting the "ignition zone" into two sections that had the same initial temperature (e.g. 2200 k for temperature, 73000 Pa for the pressure) and having -350 m/s for the left traveling pulse and 350 m/s for the right traveling pulse. I left all other sections of the pipe as nearly ambient (i.e. 300k, 10000 pa). I ended up getting:
• * A compression ration of 3/1
* A thrust of 1kg/sec from a 1/2" throat (not sure I calculated this correctly)
* A waveform that suggests that with the right tuning, you can get almost perfect out-of-phase pulses from each end of the engine
The design was a 1' x 1" main chamber with two 8" x 1/2" nozzles.

One problem that you may encounter is that the length needs to be greater than the diameter of the pipe -- took me a few crashes to figure that one out even though it's probably in the manual.

I am hoping to come up with a way to slice the pipe up into a bunch of smal segments, run a brief simulation, stop it at a given point and then use the results as ICs for the next simulation -- basically a poor-mans simuation of forced response. The idea I think that models like this will all converge to similar steady-state responses independent on initial conditions.[/list]

Bruno Ogorelec
Posts: 3542
Joined: Sat Oct 04, 2003 7:31 am
Antipspambot question: 0
Location: Zagreb, Croatia

Re: UFlow

pezman wrote:
• * A compression ration of 3/1
* A thrust of 1kg/sec from a 1/2" throat (not sure I calculated this correctly)
* A waveform that suggests that with the right tuning, you can get almost perfect out-of-phase pulses from each end of the engine
Wow! Not bad by half in terms of compression. However, the thrust figure looks low. If the thing is any improvement over the conventional designs, a half-inch nozzle should be giving over 1.5 kilos.

I sure hope you build the thing.

pezman
Posts: 613
Joined: Fri Oct 08, 2004 4:13 am
Antipspambot question: 0
Location: USA

Re: UFlow

I guess that I still have two questions:
• * Is the thrust just the mass flow rate at the nozzle?
* Is it reasonable to simulate combustion with propane as:
having a temperature of about 2200k,
with presure scaled accordingly (i.e. 730000 PA at 1:1 compresion, 2190000 at 3:1 compression etc.) and
with initial velocities of mach1 in opposite directions
using two zones that are 1 pipe diameter in length
Also, the simulation just ran just the first pulse (1:1) compression. Subsequent pulses should generate more thrust due to the higher compression. That's why I want a way to automatically feed in samples from one run as ICs in subsequent runs. For example, you could take the pressure and velocity results from a point in a given run (e.g. when the compression pulse hits the other end of the tube) and use those conditions as ICs for the resulting pulse.

I can do it by hand for now.

The engine is designed to be made from off-the-shelf plumbing parts, so it should be easy enough to build.[/list]

Bruno Ogorelec
Posts: 3542
Joined: Sat Oct 04, 2003 7:31 am
Antipspambot question: 0
Location: Zagreb, Croatia

Re: UFlow

Didn't even notice. Happens all the time. Don't worry.
pezman wrote:Is the thrust just the mass flow rate at the nozzle?
Thrust is the product of the mass flow rate and the exhaust gas velocity.
pezman wrote:Is it reasonable to simulate combustion with propane as having a temperature of about 2200k, with presure scaled accordingly (i.e. 730000 PA at 1:1 compresion, 2190000 at 3:1 compression etc.) and with initial velocities of mach1 in opposite directions using two zones that are 1 pipe diameter in length
Whoa! Too much for me. But, let me try. If I put a foot in my mouth, someone wiser will point it out and I can extract it.

The temperature looks right (maybe a bit low for the temperature of combustion but it must drop almost immediately to about what you have assumed. Pressure scaling sounds OK to me. The initial velocity is probably Mach 1 in local terms, which is what counts.

Looking forward to your engine. Don't expect it to work. Pulsejets are cranky bastards and experimental designs are much crankier than the norm.

Vaulter
Posts: 28
Joined: Fri Oct 08, 2004 5:19 pm
Antipspambot question: 0
Location: San Diego

Re: UFlow

Sorry to bother you guys, but I have been searching through the sight and have not been able to find anywhere to get the uflow software. Does anyone have some suggestions. Thanks
Last edited by Vaulter on Wed Oct 13, 2004 10:24 pm, edited 1 time in total.

Bruno Ogorelec
Posts: 3542
Joined: Sat Oct 04, 2003 7:31 am
Antipspambot question: 0
Location: Zagreb, Croatia

Re: UFlow

Vaulter wrote:Sorry to bother you guys, but I have been searching through the sight and have not been able to find anywhere to get the uflow software. Does anyone have some suggestions. Thanks

larry cottrill
Posts: 4140
Joined: Sun Oct 05, 2003 1:17 am
Antipspambot question: 0
Location: Mingo, Iowa USA
Contact:

Re: UFlow

Bruno Ogorelec wrote:Thrust is the product of the mass flow rate and the exhaust gas velocity.
This throws some people at first, since the two are closely related [i.e. if you throw the gas out faster, the flow rate will obviously be higher], but they are conceptually different.

Imagine you are sitting in a frictionless wheelchair on a basketball floor, and your job is to propel yourself across the floor by doing nothing but throwing basketballs toward the wall you're facing. An assistant will feed you basketballs at any rate you can take them. If you're athletic enough, you can throw 3 basketballs per second; that's your maximum flow rate. If you just kind of floop them on the floor in front of you, you're not going to do much, even with 3 per second. If you really heave them at the wall, you'll really sail, even though your flow rate's the same - that's the effect of increased velocity. On the other hand, if you really heave them, but decide to handle only one every ten seconds, you again won't get much - that's the effect of decreased flow rate.

Thrust in a reaction engine is really the momentum transferred per unit time, and momentum is mass x velocity. So, thrust is mass/sec x velocity.

L Cottrill

Bruno Ogorelec
Posts: 3542
Joined: Sat Oct 04, 2003 7:31 am
Antipspambot question: 0
Location: Zagreb, Croatia

Re: UFlow

Larry is right, of Course. However, the simple point (often missed) is that in practical life, thrust is indistinguishable from force. Put your engine onto scales in the vertical position and fire it upwards. What the scales show beyond the weight of the engine itself is thrust.

Graham C. Williams
Posts: 571
Joined: Fri Oct 03, 2003 11:33 pm
Antipspambot question: 0
Location: England
Contact:

Re: UFlow

Dear Pezman.

Atmospheric pressure is about 101.325KPa or 101325Pa. For most things we can use 100000Pa in UFLOW to donate atmospheric pressure.
Twice atmospheric pressure is about 200KPa and 3 times atmospheric pressure is about 300KPa. Unless you are designing a BCVP type motor you will not need to go much above 200KPa in your work.
Also, try not to confuse wave velocity with the particle velocity. Yes, the two are connected and influence each other but keep the distinction clear at least in the early phases of your work, unless you are planning a PDE.

Graham.

Bruno Ogorelec
Posts: 3542
Joined: Sat Oct 04, 2003 7:31 am
Antipspambot question: 0
Location: Zagreb, Croatia

Re: UFlow

Graham, I think our friend was talking of the compression ratio, rather than the pressure peak. In the conventional pulsejet, I agree that the peak pressure is probably not much above 200kPa but if the pre-compression ratio is 1:3, the pressure at the start of combustion will already be 300kPa and will climb from there due to generation of combustion products.

Graham C. Williams
Posts: 571
Joined: Fri Oct 03, 2003 11:33 pm
Antipspambot question: 0
Location: England
Contact:

Re: UFlow

Dear Bruno.
Ok, but I'm still unshure how you generate these internal pressures with a pressure gradient of only 0.5bar across the motor?

Good night my friend.
Graham.

Bruno Ogorelec
Posts: 3542
Joined: Sat Oct 04, 2003 7:31 am
Antipspambot question: 0
Location: Zagreb, Croatia

Re: UFlow

Ah, this is a very good question, Graham. It is so easy to arrive at high pressures in your head... In my head, long ago, I was secretly hoping for this:

First bang, first chamber
Pressure rises from atmospheric (1 At) to twice that because of combustion.

Second bang, second chamber
Precompression up to 1:2 because of blast compression. Combustion ups this to 4 At.

Third bang, first chamber
Precompression 1:4. Pressure peak 8 At.

And so it went.

Two processes would moderate the pressure rise. One is the fact that a certain amount of gas is diverted to generate thrust. The other is that the pressure rise would at some point start limiting the amount of fresh charge the engine would be able to ingest.

Frankly, I was hoping the thing might stabilize somewhere at perhaps 500 kPa maximum. Not that I had any strong reasonfor this; it was just a modest hope.

pezman
Posts: 613
Joined: Fri Oct 08, 2004 4:13 am
Antipspambot question: 0
Location: USA

Re: UFlow

Two processes would moderate the pressure rise. One is the fact that a certain amount of gas is diverted to generate thrust. The other is that the pressure rise would at some point start limiting the amount of fresh charge the engine would be able to ingest.
UFlow agrees! The "Plumbing" design (just a straight pipe w/ restrictions at the end) did not seem to work because the mean internal pressure of the device would eventually cause it to stop drawing air. That motivated me to build the simple lockwood model, to ensure that UFlow would properly model the breathing of the engine (which it did), and that led to the idea of gluing two half-lockwoods together -- which seems to have some promise.

The geometry of the pipe between the two chambers looks to be important. The only model that I've come up with so far that works well is just two tapered tubes with the narrow ends welded together and short nozzles sticking out of each chamber (e.g. -><- )[/i][/quote]